TWI677843B - Intelligent cluster suggestion system and method - Google Patents

Intelligent cluster suggestion system and method Download PDF

Info

Publication number
TWI677843B
TWI677843B TW106131682A TW106131682A TWI677843B TW I677843 B TWI677843 B TW I677843B TW 106131682 A TW106131682 A TW 106131682A TW 106131682 A TW106131682 A TW 106131682A TW I677843 B TWI677843 B TW I677843B
Authority
TW
Taiwan
Prior art keywords
cluster
server unit
transaction
transaction characteristic
data
Prior art date
Application number
TW106131682A
Other languages
Chinese (zh)
Other versions
TW201915899A (en
Inventor
發明人放棄姓名表示權
Inventor Has Agreed To Waive The Entitlement To Designation This
Original Assignee
群益金鼎證券股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 群益金鼎證券股份有限公司 filed Critical 群益金鼎證券股份有限公司
Priority to TW106131682A priority Critical patent/TWI677843B/en
Publication of TW201915899A publication Critical patent/TW201915899A/en
Application granted granted Critical
Publication of TWI677843B publication Critical patent/TWI677843B/en

Links

Landscapes

  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一種智能分群分析系統,包含一伺服器單元及一終端電子裝置。該伺服器單元根據多個客戶交易記錄產生多個交易特徵資料,每一交易特徵資料包含多個交易特徵值(其中一者為一總損益)。該伺服器單元根據該等交易特徵資料產生多個群集定義資料並將該等交易特徵資料劃分為多個群集。該伺服器單元針對每一群集,於該群集的該等交易特徵資料中,選定該總損益在該群集中的百分位符合一預定百分位範圍之該等交易特徵資料分別做為多個目標交易特徵資料。該伺服器單元針對每一群集,根據該群集的該等目標交易特徵資料的該等交易特徵值產生一參考資料。An intelligent cluster analysis system includes a server unit and a terminal electronic device. The server unit generates multiple transaction characteristic data according to multiple customer transaction records, and each transaction characteristic data includes multiple transaction characteristic values (one of which is a total profit or loss). The server unit generates a plurality of cluster definition data based on the transaction characteristic data and divides the transaction characteristic data into a plurality of clusters. For each cluster, the server unit selects, among the transaction characteristic data of the cluster, the transaction characteristic information of which the percentile of the total profit or loss in the cluster meets a predetermined percentile range, respectively. Target transaction characteristics. For each cluster, the server unit generates a reference data according to the transaction characteristic values of the target transaction characteristic data of the cluster.

Description

智能分群建議系統及方法Intelligent cluster suggestion system and method

本發明是有關於一種智能分析系統,特別是指一種使用分群技術的智能分析系統。本發明還有關於一種使用分群技術的智能分析方法。The invention relates to an intelligent analysis system, in particular to an intelligent analysis system using a clustering technology. The invention also relates to an intelligent analysis method using clustering technology.

隨著網路通訊及電腦技術的蓬勃發展,投資者可以使用電腦或智慧型手機即時地完成金融商品(例如股票)交易。如何適當地應用交易系統中眾多投資者的交易記錄,配合大數據分析技術,提供投資者個人的投資行為分析及建議,為本案進一步要探討的主題。With the rapid development of Internet communications and computer technology, investors can use computers or smartphones to complete transactions in financial commodities (such as stocks) in real time. How to properly apply the transaction records of many investors in the trading system and cooperate with big data analysis technology to provide individual investors' investment behavior analysis and recommendations is the subject of this case.

因此,本發明的目的,即在提供一種智能分群分析系統。Therefore, an object of the present invention is to provide an intelligent cluster analysis system.

本發明的另一目的,在於提供一種智能分群分析方法。Another object of the present invention is to provide an intelligent cluster analysis method.

於是,本發明智能分群分析系統,包含一伺服器單元及一終端電子裝置。該伺服器單元儲存有分別相關於多個客戶的客戶交易記錄。該終端電子裝置能與該伺服器單元通訊。Therefore, the intelligent cluster analysis system of the present invention includes a server unit and a terminal electronic device. The server unit stores customer transaction records related to a plurality of customers, respectively. The terminal electronic device can communicate with the server unit.

該伺服器單元根據該等客戶交易記錄產生多個分別相關於該等客戶的交易特徵資料,該等交易特徵資料的每一者包含多個交易特徵值,該等交易特徵值的其中一者為一總損益。The server unit generates a plurality of transaction characteristic data related to the customers according to the transaction records of the customers. Each of the transaction characteristic data includes a plurality of transaction characteristic values. One of the transaction characteristic values is A total profit and loss.

該伺服器單元根據該等交易特徵資料,使用一預定分群演算法,產生多個群集定義資料,該等群集定義資料的每一者包含多個分別相關於該等交易特徵值的特徵值範圍。The server unit generates a plurality of cluster definition data using a predetermined grouping algorithm based on the transaction characteristic data, and each of the cluster definition data includes a plurality of characteristic value ranges respectively related to the transaction characteristic values.

該伺服器單元根據該等群集定義資料將該等交易特徵資料劃分為多個分別對應於該等群集定義資料的群集。The server unit divides the transaction characteristic data into a plurality of clusters respectively corresponding to the cluster definition data according to the cluster definition data.

該伺服器單元針對該等群集的每一者,於該群集的該等交易特徵資料中,選定該總損益在該群集中的百分位符合一預定百分位範圍之該等交易特徵資料分別做為多個目標交易特徵資料。For each of the clusters, the server unit selects, among the transaction characteristic data of the cluster, the transaction characteristic information of which the percentile of the total profit or loss in the cluster meets a predetermined percentile range, respectively As multiple target transaction characteristic data.

該伺服器單元針對該等群集的每一者,根據該群集的該等目標交易特徵資料的該等交易特徵值,產生一對應於該群集的參考資料,該參考資料包含多個分別相關於該等交易特徵值的參考特徵值。The server unit generates, for each of the clusters, a reference material corresponding to the cluster according to the transaction characteristic values of the target transaction characteristic data of the cluster, and the reference material includes a plurality of references related to the cluster. The reference characteristic value of the characteristic value of the transaction.

該終端電子裝置傳送一相關於該等客戶其中一目標者的分析請求給該伺服器單元。The terminal electronic device sends an analysis request related to one of the clients to the server unit.

當該伺服器單元接收到該分析請求,該伺服器單元傳送一分析結果給該終端電子裝置,該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值,及相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值。When the server unit receives the analysis request, the server unit transmits an analysis result to the terminal electronic device, and the analysis result includes at least part of the transaction characteristic values related to the transaction characteristic data of the target customer, and the related At least part of the reference characteristic values of the reference data corresponding to the cluster to which the transaction characteristic data of the target customer belongs.

當該終端電子裝置接收到該分析結果,該終端電子裝置顯示該分析結果。When the terminal electronic device receives the analysis result, the terminal electronic device displays the analysis result.

在一些實施態樣中,該伺服器單元根據該等特徵值範圍,使用一預定決策樹演算法,決定多個分別相關於該等群集定義資料的群集屬性描述。該分析結果還包含相關於該目標客戶的該交易特徵資料所符合之該群集定義資料所相關之該群集屬性描述。In some implementations, the server unit determines a plurality of cluster attribute descriptions respectively related to the cluster definition data using a predetermined decision tree algorithm based on the feature value ranges. The analysis result also includes the cluster attribute description related to the cluster definition data corresponding to the transaction characteristic data of the target customer.

在一些實施態樣中,該伺服器單元還儲存有多個分別相關於多個金融商品的金融商品歷史資料。該伺服器單元根據該等金融商品歷史資料產生多個分別相關於該等金融商品的商品特徵資料,該等商品特徵資料的每一者包含多個商品特徵值。該終端電子裝置傳送一相關於該目標客戶的建議請求給該伺服器單元,該建議請求包含至少一相關於該分析結果的該等交易特徵值其中一者之目標特徵值。當該伺服器單元接收到該建議請求,該伺服器單元根據該等商品特徵資料、該至少一目標特徵值及相關於該目標客戶的交易特徵資料,產生一相關於該等金融商品其中至少一者的商品清單,並將該商品清單傳送給該終端電子裝置。當該終端電子裝置接收到該商品清單,該終端電子裝置顯示該商品清單。In some implementation aspects, the server unit further stores a plurality of financial commodity historical data respectively related to a plurality of financial commodities. The server unit generates a plurality of product characteristic data respectively related to the financial products based on the historical data of the financial products, and each of the product characteristic data includes a plurality of product characteristic values. The terminal electronic device sends a recommendation request related to the target customer to the server unit, the recommendation request including at least one target characteristic value related to one of the transaction characteristic values of the analysis result. When the server unit receives the suggestion request, the server unit generates at least one of the financial commodities related to the financial commodities according to the commodity characteristic data, the at least one target characteristic value, and the transaction characteristic data related to the target customer. The product list of the user, and transmits the product list to the terminal electronic device. When the terminal electronic device receives the product list, the terminal electronic device displays the product list.

在一些實施態樣中,該預定分群演算法為k-平均演算法。In some implementation aspects, the predetermined group algorithm is a k-average algorithm.

在一些實施態樣中,該終端電子裝置是以雷達圖的形式顯示該分析結果。In some embodiments, the terminal electronic device displays the analysis result in the form of a radar chart.

本發明智能分群分析方法,藉由一智能分群分析系統實施,該智能分群分析系統包含一伺服器單元及一終端電子裝置,該伺服器單元儲存有分別相關於多個客戶的客戶交易記錄,該終端電子裝置能與該伺服器單元通訊,該方法包含:該伺服器單元根據該等客戶交易記錄產生多個分別相關於該等客戶的交易特徵資料,該等交易特徵資料的每一者包含多個交易特徵值,該等交易特徵值的其中一者為一總損益;該伺服器單元根據該等交易特徵資料,使用一預定分群演算法,產生多個群集定義資料,該等群集定義資料的每一者包含多個分別相關於該等交易特徵值的特徵值範圍;該伺服器單元根據該等群集定義資料將該等交易特徵資料劃分為多個分別對應於該等群集定義資料的群集;該伺服器單元針對該等群集的每一者,於該群集的該等交易特徵資料中,選定該總損益在該群集中的百分位符合一預定百分位範圍之該等交易特徵資料分別做為多個目標交易特徵資料;該伺服器單元針對該等群集的每一者,根據該群集的該等目標交易特徵資料的該等交易特徵值,產生一對應於該群集的參考資料,該參考資料包含多個分別相關於該等交易特徵值的參考特徵值;該終端電子裝置傳送一相關於該等客戶其中一目標者的分析請求給該伺服器單元;當該伺服器單元接收到該分析請求,該伺服器單元傳送一分析結果給該終端電子裝置,該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值,及相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值;及當該終端電子裝置接收到該分析結果,該終端電子裝置顯示該分析結果。The intelligent cluster analysis method of the present invention is implemented by an intelligent cluster analysis system. The intelligent cluster analysis system includes a server unit and a terminal electronic device. The server unit stores customer transaction records related to multiple customers. The terminal electronic device can communicate with the server unit, and the method includes: the server unit generates a plurality of transaction characteristic data respectively related to the customers according to the transaction records of the customers, and each of the transaction characteristic data includes multiple One transaction characteristic value, one of the transaction characteristic values is a total profit and loss; the server unit generates a plurality of cluster definition data based on the transaction characteristic data using a predetermined group algorithm, Each of which includes a plurality of characteristic value ranges respectively related to the transaction characteristic values; the server unit divides the transaction characteristic data into a plurality of clusters respectively corresponding to the cluster definition data according to the cluster definition data; The server unit selects, for each of the clusters, among the transaction characteristic data of the cluster. The transaction characteristic data of which the percentile of the total profit or loss in the cluster meets a predetermined percentile range is respectively regarded as a plurality of target transaction characteristic data; the server unit targets each of the clusters according to the cluster The transaction characteristic values of the target transaction characteristic data of the generated a reference material corresponding to the cluster, and the reference material contains a plurality of reference characteristic values respectively related to the transaction characteristic values; the terminal electronic device transmits a correlation An analysis request from one of the customers of the target is sent to the server unit; when the server unit receives the analysis request, the server unit sends an analysis result to the terminal electronic device, and the analysis result includes information related to the target At least part of the transaction characteristic values of the transaction characteristic information of the customer, and at least part of the reference characteristic values of the reference information corresponding to the cluster to which the transaction characteristic information of the target customer belongs; and when the terminal electronics The device receives the analysis result, and the terminal electronic device displays the analysis result.

本發明的功效在於:藉由該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值,及相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值,從而讓該目標客戶能比較該等交易特徵值與該等參考特徵值的差異,做為調整本身投資行為的參考,此外,藉由該伺服器單元回應於該建議請求產生該商品清單,從而讓該目標客戶可以參考該商品清單調整本身投資行為,以使該目標客戶的投資行為更接近相同群集中總損益較佳的客戶。The effect of the present invention is that the analysis result includes at least part of the transaction characteristic values related to the transaction characteristic data of the target customer, and the transaction corresponding to the cluster to which the transaction characteristic data related to the target customer belongs. At least part of the reference characteristic values of the reference materials, so that the target customer can compare the difference between the transaction characteristic values and the reference characteristic values as a reference for adjusting its own investment behavior. In addition, the server unit responds The suggestion request generates the product list, so that the target customer can refer to the product list to adjust its investment behavior, so that the target customer's investment behavior is closer to the customer with better total profit and loss in the same cluster.

參閱圖1,本發明智能分群分析系統100的一實施例,包含一伺服器單元1及一終端電子裝置2。Referring to FIG. 1, an embodiment of the intelligent cluster analysis system 100 of the present invention includes a server unit 1 and a terminal electronic device 2.

該伺服器單元1儲存有分別相關於多個客戶的客戶交易記錄D1(例如客戶買賣股票之記錄),及多個分別相關於多個金融商品(例如股票)的金融商品歷史資料D2(例如除權息資料、在外流通股數等)。The server unit 1 stores customer transaction records D1 (such as records of customers buying and selling stocks) that are related to multiple customers, and a plurality of financial commodity historical data D2 (such as ex-rights) that are respectively related to multiple financial products (such as stocks). Information, number of shares outstanding, etc.).

該終端電子裝置2能經由一通訊網路與該伺服器單元1通訊。該終端電子裝置2例如是一智慧型手機、一平板電腦、一桌上型電腦或一膝上型電腦,但不以此為限。The terminal electronic device 2 can communicate with the server unit 1 via a communication network. The terminal electronic device 2 is, for example, a smart phone, a tablet computer, a desktop computer, or a laptop computer, but is not limited thereto.

以下配合圖1及圖2說明該智能分群分析系統100執行一資料分析程序的步驟。首先,如步驟S01所示,該伺服器單元1根據該等客戶交易記錄D1產生多個分別相關於該等客戶的交易特徵資料,並根據該等金融商品歷史資料D2產生多個分別相關於該等金融商品的商品特徵資料。The steps of the intelligent cluster analysis system 100 executing a data analysis program are described below with reference to FIGS. 1 and 2. First, as shown in step S01, the server unit 1 generates a plurality of transaction characteristic data related to the customers according to the customer transaction records D1, and generates a plurality of data related to the financial commodity historical data D2, respectively. Product characteristics of other financial commodities.

該等交易特徵資料的每一者包含多個交易特徵值。該等交易特徵值例如為一總損益、一投資勝率、一風險承受度、一投資頻率、一資產能力、一投機程度、一總交易天數及一總交易金額,但該等交易特徵值的項目及數目不以此為限。Each of the transaction characteristic data includes a plurality of transaction characteristic values. These transaction characteristic values are, for example, a total profit and loss, an investment win ratio, a risk tolerance, an investment frequency, an asset capacity, a speculative level, a total transaction days and a total transaction amount, but the items of these transaction characteristic values And the number is not limited to this.

該等商品特徵資料的每一者包含多個商品特徵值。該等商品特徵值例如為一風險程度、一流動率、一波動率、一營收、一籌碼、一殖利率及一技術,但該等商品特徵值的項目及數目不以此為限。Each of the commodity characteristic data includes a plurality of commodity characteristic values. The characteristic values of these commodities are, for example, a degree of risk, a liquidity rate, a volatility rate, a revenue, a chip, a yield rate, and a technology, but the items and number of the characteristic values of these commodities are not limited thereto.

接著,如步驟S02所示,該伺服器單元1根據該等交易特徵資料,使用一預定分群演算法,產生多個群集定義資料。該等群集定義資料的每一者包含多個分別相關於該等交易特徵值的特徵值範圍。在本實施例中,該預定分群演算法為k-平均演算法(k-means clustering)。Next, as shown in step S02, the server unit 1 generates a plurality of cluster definition data using a predetermined group algorithm based on the transaction characteristic data. Each of the cluster definition data includes a plurality of characteristic value ranges that are respectively related to the characteristic values of the transactions. In this embodiment, the predetermined clustering algorithm is k-means clustering.

接著,如步驟S03所示,該伺服器單元1根據該等群集定義資料將該等交易特徵資料劃分為多個分別對應於該等群集定義資料的群集。Next, as shown in step S03, the server unit 1 divides the transaction characteristic data into a plurality of clusters respectively corresponding to the cluster definition data according to the cluster definition data.

接著,如步驟S04所示,該伺服器單元1根據該等特徵值範圍,使用一預定決策樹演算法,決定多個分別相關於該等群集定義資料的群集屬性描述。該等群集屬性描述例如為「高風險型」、「養股型」、「穩健經營型」等,但不以此為限。Then, as shown in step S04, the server unit 1 determines a plurality of cluster attribute descriptions respectively related to the cluster definition data by using a predetermined decision tree algorithm according to the feature value ranges. These cluster attribute descriptions are, for example, "high risk type", "stock raising type", "stable operating type", etc., but are not limited thereto.

接著,如步驟S05所示,該伺服器單元1針對該等群集的每一者,於該群集的該等交易特徵資料中,選定該總損益在該群集中的百分位符合一預定百分位範圍之該等交易特徵資料分別做為多個目標交易特徵資料。在本實施例中,該預定百分位範圍例如為第50~第100,但不以此為限。Next, as shown in step S05, the server unit 1 selects, for each of the clusters, among the transaction characteristic data of the cluster, the percentile of the total profit or loss in the cluster to meet a predetermined percentage. These range of transaction characteristic data are used as multiple target transaction characteristic data, respectively. In this embodiment, the predetermined percentile range is, for example, 50th to 100th, but is not limited thereto.

最後,如步驟S06所示,該伺服器單元1針對該等群集的每一者,根據該群集的該等目標交易特徵資料的該等交易特徵值,產生一對應於該群集的參考資料,該參考資料包含多個分別相關於該等交易特徵值的參考特徵值。Finally, as shown in step S06, the server unit 1 generates, for each of the clusters, a reference material corresponding to the cluster according to the transaction characteristic values of the target transaction characteristic data of the cluster. The reference material contains a plurality of reference characteristic values that are respectively related to the transaction characteristic values.

該智能分群分析系統100可以是每隔預定時間執行該資料分析程序以更新分析結果。The intelligent cluster analysis system 100 may execute the data analysis program every predetermined time to update the analysis results.

以下配合圖1及圖3說明該智能分群分析系統100執行一資料回饋程序的步驟。首先,如步驟S11所示,該終端電子裝置2根據該目標客戶的操作,傳送一相關於該等客戶其中一目標者的分析請求給該伺服器單元1。The steps of the intelligent cluster analysis system 100 executing a data feedback procedure are described below with reference to FIGS. 1 and 3. First, as shown in step S11, the terminal electronic device 2 transmits an analysis request related to one of the target customers to the server unit 1 according to the operation of the target customer.

接著,如步驟S12所示,當該伺服器單元1接收到該分析請求,該伺服器單元1傳送一分析結果給該終端電子裝置2。該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值(在本實施例中是該投資勝率、該風險承受度、該投資頻率、該資產能力及該投機程度共五項交易特徵值)、相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值(在本實施例中是該投資勝率、該風險承受度、該投資頻率、該資產能力及該投機程度共五項參考特徵值),及相關於該目標客戶的該交易特徵資料所符合之該群集定義資料所相關之該群集屬性描述。Next, as shown in step S12, when the server unit 1 receives the analysis request, the server unit 1 transmits an analysis result to the terminal electronic device 2. The analysis result includes at least part of the transaction characteristic data related to the target customer's transaction characteristic data (in this embodiment, the investment win rate, the risk tolerance, the investment frequency, the asset capacity and the speculative degree are total). Five transaction characteristic values), at least part of the reference characteristic values (in this embodiment, the investment win rate, the risk tolerance) of the reference material corresponding to the cluster to which the transaction characteristic data of the target customer belongs , The investment frequency, the asset capacity, and the degree of speculation, a total of five reference characteristic values), and the cluster attribute description related to the cluster definition data related to the transaction characteristic data of the target customer.

接著,如步驟S13所示,當該終端電子裝置2接收到該分析結果,該終端電子裝置2顯示該分析結果。如圖4所示,該終端電子裝置2是以雷達圖的形式顯示該分析結果,其中,該雷達圖呈現本實施例中相關於該目標客戶的該交易特徵資料的五項交易特徵值(該投資勝率、該風險承受度、該投資頻率、該資產能力及該投機程度)分別是20、29、43、54及26,且該雷達圖呈現本實施例中相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的五項參考特徵值(該投資勝率、該風險承受度、該投資頻率、該資產能力及該投機程度)分別是70、52、43、54及73。藉此,該目標客戶能比較該等交易特徵值與該等參考特徵值的差異,做為調整本身投資行為的參考。Next, as shown in step S13, when the terminal electronic device 2 receives the analysis result, the terminal electronic device 2 displays the analysis result. As shown in FIG. 4, the terminal electronic device 2 displays the analysis result in the form of a radar chart, where the radar chart presents five transaction characteristic values (the The investment win rate, the risk tolerance, the investment frequency, the asset capacity and the degree of speculation are 20, 29, 43, 54 and 26 respectively, and the radar chart presents the transaction characteristics related to the target customer in this embodiment The five reference characteristic values (the investment win rate, the risk tolerance, the investment frequency, the asset capacity, and the degree of speculation) of the reference material corresponding to the cluster to which the data belongs are 70, 52, 43, 54 and 73, respectively . In this way, the target customer can compare the differences between the transaction characteristic values and the reference characteristic values as a reference for adjusting its investment behavior.

接著,如步驟S14所示,該終端電子裝置2根據該目標客戶的操作,傳送一相關於該目標客戶的建議請求給該伺服器單元1,該建議請求包含至少一相關於該分析結果的該等交易特徵值其中一者之目標特徵值。如圖5所示,該目標客戶可以透過操作該終端電子裝置2將該投資勝率的值由20的位置拖曳至70的位置,以將該目標特徵值設定為70。Then, as shown in step S14, the terminal electronic device 2 transmits a recommendation request related to the target customer to the server unit 1 according to the operation of the target customer, and the recommendation request includes at least one related to the analysis result. The target characteristic value of one of the transaction characteristic values. As shown in FIG. 5, the target customer can drag the value of the investment winning ratio from the position of 20 to the position of 70 by operating the terminal electronic device 2 to set the target characteristic value to 70.

接著,如步驟S15所示,當該伺服器單元1接收到該建議請求,該伺服器單元1根據該等商品特徵資料、該至少一目標特徵值及相關於該目標客戶的該交易特徵資料,產生一相關於該等金融商品其中至少一者的商品清單。Then, as shown in step S15, when the server unit 1 receives the suggestion request, the server unit 1 according to the product characteristic data, the at least one target characteristic value, and the transaction characteristic data related to the target customer, A commodity list is generated for at least one of the financial commodities.

接著,如步驟S16所示,該伺服器單元1將該商品清單傳送給該終端電子裝置2。Then, as shown in step S16, the server unit 1 transmits the product list to the terminal electronic device 2.

最後,如步驟S17所示,當該終端電子裝置2接收到該商品清單,該終端電子裝置2顯示該商品清單。若該目標客戶參考該商品清單而購買該商品清單所相關的該至少一金融商品,則相關於該目標客戶的該交易特徵資料的其中至少一交易特徵值將會趨近該至少一目標特徵值(例如該投資勝率會由20改變為趨近於70)。藉此,該目標客戶可以參考該商品清單調整本身投資行為,以使該目標客戶的投資行為更接近相同群集中總損益較佳的客戶。Finally, as shown in step S17, when the terminal electronic device 2 receives the product list, the terminal electronic device 2 displays the product list. If the target customer purchases the at least one financial commodity related to the product list with reference to the product list, at least one transaction characteristic value of the transaction characteristic data related to the target customer will approach the at least one target characteristic value (For example, the investment win ratio will change from 20 to close to 70). In this way, the target customer can adjust its investment behavior by referring to the product list, so that the target customer's investment behavior is closer to the customer with better total profit and loss in the same cluster.

綜上所述,本發明智能分群分析系統100的實施例,藉由該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值,及相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值,從而讓該目標客戶能比較該等交易特徵值與該等參考特徵值的差異,做為調整本身投資行為的參考,此外,藉由該伺服器單元1回應於該建議請求而產生該商品清單,從而讓該目標客戶可以參考該商品清單調整本身投資行為,以使該目標客戶的投資行為更接近相同群集中總損益較佳的客戶,故確實能達成本發明的目的。In summary, in the embodiment of the intelligent cluster analysis system 100 of the present invention, the analysis result includes at least part of the transaction characteristic values related to the transaction characteristic data of the target customer, and the transaction related to the target customer. At least part of the reference characteristic values of the reference material corresponding to the cluster to which the characteristic data belongs, so that the target customer can compare the difference between the transaction characteristic values and the reference characteristic values as a reference for adjusting its own investment behavior In addition, the server unit 1 generates the product list in response to the suggestion request, so that the target customer can refer to the product list to adjust its own investment behavior so that the target customer's investment behavior is closer to the total in the same cluster. Customers with better profit and loss can indeed achieve the purpose of cost invention.

惟以上所述者,僅為本發明的實施例而已,當不能以此限定本發明實施的範圍,凡是依本發明申請專利範圍及專利說明書內容所作的簡單的等效變化與修飾,皆仍屬本發明專利涵蓋的範圍內。However, the above are only examples of the present invention. When the scope of implementation of the present invention cannot be limited by this, any simple equivalent changes and modifications made according to the scope of the patent application and the contents of the patent specification of the present invention are still Within the scope of the invention patent.

100‧‧‧智能分群分析系統100‧‧‧ Intelligent Cluster Analysis System

1‧‧‧伺服器單元1‧‧‧Server Unit

D1‧‧‧客戶交易記錄D1‧‧‧Customer transaction records

D2‧‧‧金融商品歷史資料D2‧‧‧Financial Commodities Historical Data

2‧‧‧終端電子裝置2‧‧‧ terminal electronics

S01~S06‧‧‧步驟S01 ~ S06‧‧‧step

S01~S17‧‧‧步驟 S01 ~ S17‧‧‧step

本發明的其他的特徵及功效,將於參照圖式的實施方式中清楚地呈現,其中: 圖1是本發明智能分群分析系統的一實施例的一硬體連接關係示意圖; 圖2是該實施例的一流程圖,說明一資料分析程序的步驟; 圖3是該實施例的另一流程圖,說明一資料回饋程序的步驟; 及 圖4與圖5分別是該實施例的一雷達圖。Other features and effects of the present invention will be clearly presented in the embodiment with reference to the drawings, wherein: FIG. 1 is a schematic diagram of a hardware connection relationship of an embodiment of the intelligent cluster analysis system of the present invention; FIG. 2 is the implementation A flowchart of the example illustrates the steps of a data analysis program; FIG. 3 is another flowchart of the embodiment, illustrating the steps of a data feedback program; and FIG. 4 and FIG. 5 are a radar chart of the embodiment, respectively.

Claims (8)

一種智能分群分析系統,包含:一伺服器單元,儲存有分別相關於多個客戶的客戶交易記錄;及一終端電子裝置,能與該伺服器單元通訊;該伺服器單元根據該等客戶交易記錄產生多個分別相關於該等客戶的交易特徵資料,該等交易特徵資料的每一者包含多個交易特徵值,該等交易特徵值的其中一者為一總損益;該伺服器單元根據該等交易特徵資料,使用一預定分群演算法,產生多個群集定義資料,該等群集定義資料的每一者包含多個分別相關於該等交易特徵值的特徵值範圍;該伺服器單元根據該等群集定義資料將該等交易特徵資料劃分為多個分別對應於該等群集定義資料的群集;該伺服器單元根據該等特徵值範圍,使用一預定決策樹演算法,決定多個分別相關於該等群集定義資料的群集屬性描述;該伺服器單元針對該等群集的每一者,於該群集的該等交易特徵資料中,選定該總損益在該群集中的百分位符合一預定百分位範圍之該等交易特徵資料分別做為多個目標交易特徵資料;該伺服器單元針對該等群集的每一者,根據該群集的該等目標交易特徵資料的該等交易特徵值,產生一對應於該群集的參考資料,該參考資料包含多個分別相關於該等交易特徵值的參考特徵值;該終端電子裝置傳送一相關於該等客戶其中一目標者的分析請求給該伺服器單元;當該伺服器單元接收到該分析請求,該伺服器單元傳送一分析結果給該終端電子裝置,該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值、相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值,及相關於該目標客戶的該交易特徵資料所符合之該群集定義資料所相關之該群集屬性描述;當該終端電子裝置接收到該分析結果,該終端電子裝置顯示該分析結果。An intelligent cluster analysis system includes: a server unit storing customer transaction records related to multiple customers; and a terminal electronic device capable of communicating with the server unit; the server unit is based on the customer transaction records Generating a plurality of transaction characteristic data related to the customers, each of which includes a plurality of transaction characteristic values, and one of the transaction characteristic values is a total profit and loss; the server unit according to the And other transaction characteristic data, using a predetermined grouping algorithm to generate a plurality of cluster definition data, each of the cluster definition data includes a plurality of characteristic value ranges respectively related to the transaction characteristic values; the server unit according to the Etc. The cluster characteristic data is divided into a plurality of clusters corresponding to the cluster definition data; the server unit uses a predetermined decision tree algorithm to determine a plurality of A cluster attribute description of the cluster definition data; the server unit for each of the clusters in the cluster Among the transaction characteristic data, the transaction characteristic data of which the percentile of the total profit or loss in the cluster meets a predetermined percentile range is selected as a plurality of target transaction characteristic data; the server unit targets the clusters. Each of which generates a reference material corresponding to the cluster according to the transaction characteristic values of the target transaction characteristic data of the cluster, and the reference material includes a plurality of reference characteristic values respectively related to the transaction characteristic values ; The terminal electronic device sends an analysis request related to one of the clients to the server unit; when the server unit receives the analysis request, the server unit sends an analysis result to the terminal electronic device, The analysis result includes at least a portion of the transaction characteristic values of the transaction characteristic data related to the target customer, and at least a portion of the reference characteristics of the reference data corresponding to the cluster to which the transaction characteristic data related to the target customer belongs. Value, and the cluster to which the cluster definition data associated with the transaction characteristic data related to the target customer corresponds Attribute Description; terminal when the electronic device receives the analysis result, the terminal electronic device displays the analysis result. 如請求項1所述的智能分群分析系統,其中,該伺服器單元還儲存有多個分別相關於多個金融商品的金融商品歷史資料;該伺服器單元根據該等金融商品歷史資料產生多個分別相關於該等金融商品的商品特徵資料,該等商品特徵資料的每一者包含多個商品特徵值;該終端電子裝置傳送一相關於該目標客戶的建議請求給該伺服器單元,該建議請求包含至少一相關於該分析結果的該等交易特徵值其中一者之目標特徵值;當該伺服器單元接收到該建議請求,該伺服器單元根據該等商品特徵資料、該至少一目標特徵值及相關於該目標客戶的交易特徵資料,產生一相關於該等金融商品其中至少一者的商品清單,並將該商品清單傳送給該終端電子裝置;當該終端電子裝置接收到該商品清單,該終端電子裝置顯示該商品清單。The intelligent cluster analysis system according to claim 1, wherein the server unit further stores a plurality of financial commodity historical data respectively related to a plurality of financial commodities; the server unit generates a plurality of financial commodity historical data according to the financial commodity historical data Product characteristic data related to the financial commodities, each of which includes a plurality of product characteristic values; the terminal electronic device sends a request for a proposal related to the target customer to the server unit, the proposal The request includes at least one target characteristic value of one of the transaction characteristic values related to the analysis result; when the server unit receives the suggestion request, the server unit according to the product characteristic data, the at least one target characteristic Value and transaction characteristic data related to the target customer, generate a product list related to at least one of the financial products, and transmit the product list to the terminal electronic device; when the terminal electronic device receives the product list , The terminal electronic device displays the product list. 如請求項1所述的智能分群分析系統,其中,該預定分群演算法為k-平均演算法。The intelligent cluster analysis system according to claim 1, wherein the predetermined cluster algorithm is a k-average algorithm. 如請求項1所述的智能分群分析系統,其中,該終端電子裝置是以雷達圖的形式顯示該分析結果。The intelligent cluster analysis system according to claim 1, wherein the terminal electronic device displays the analysis result in the form of a radar chart. 一種智能分群分析方法,藉由一智能分群分析系統實施,該智能分群分析系統包含一伺服器單元及一終端電子裝置,該伺服器單元儲存有分別相關於多個客戶的客戶交易記錄,該終端電子裝置能與該伺服器單元通訊,該方法包含:該伺服器單元根據該等客戶交易記錄產生多個分別相關於該等客戶的交易特徵資料,該等交易特徵資料的每一者包含多個交易特徵值,該等交易特徵值的其中一者為一總損益;該伺服器單元根據該等交易特徵資料,使用一預定分群演算法,產生多個群集定義資料,該等群集定義資料的每一者包含多個分別相關於該等交易特徵值的特徵值範圍;該伺服器單元根據該等群集定義資料將該等交易特徵資料劃分為多個分別對應於該等群集定義資料的群集;該伺服器單元根據該等特徵值範圍,使用一預定決策樹演算法,決定多個分別相關於該等群集定義資料的群集屬性描述;該伺服器單元針對該等群集的每一者,於該群集的該等交易特徵資料中,選定該總損益在該群集中的百分位符合一預定百分位範圍之該等交易特徵資料分別做為多個目標交易特徵資料;該伺服器單元針對該等群集的每一者,根據該群集的該等目標交易特徵資料的該等交易特徵值,產生一對應於該群集的參考資料,該參考資料包含多個分別相關於該等交易特徵值的參考特徵值;該終端電子裝置傳送一相關於該等客戶其中一目標者的分析請求給該伺服器單元;當該伺服器單元接收到該分析請求,該伺服器單元傳送一分析結果給該終端電子裝置,該分析結果包含相關於該目標客戶的該交易特徵資料的至少部分該等交易特徵值、相關於該目標客戶的該交易特徵資料所屬之該群集所對應之該參考資料的至少部分該等參考特徵值,及相關於該目標客戶的該交易特徵資料所符合之該群集定義資料所相關之該群集屬性描述;及當該終端電子裝置接收到該分析結果,該終端電子裝置顯示該分析結果。An intelligent clustering analysis method is implemented by an intelligent clustering analysis system. The intelligent clustering analysis system includes a server unit and a terminal electronic device. The server unit stores customer transaction records related to multiple customers. The terminal The electronic device can communicate with the server unit. The method includes: the server unit generates a plurality of transaction characteristic data related to the customers according to the customer transaction records, and each of the transaction characteristic data includes a plurality of Transaction characteristic value, one of the transaction characteristic values is a total profit and loss; the server unit generates a plurality of cluster definition data based on the transaction characteristic data using a predetermined group algorithm, and each of the cluster definition data One includes a plurality of characteristic value ranges respectively related to the transaction characteristic values; the server unit divides the transaction characteristic data into a plurality of clusters respectively corresponding to the cluster definition data according to the cluster definition data; the The server unit uses a predetermined decision tree algorithm to determine multiple A description of the cluster properties of the cluster definition data; the server unit, for each of the clusters, selects, in the transaction characteristic data of the cluster, the percentile of the total profit or loss in the cluster meets a predetermined The transaction characteristic data in the percentile range are respectively used as a plurality of target transaction characteristic data; the server unit targets each of the clusters according to the transaction characteristic values of the target transaction characteristic data of the cluster, Generate a reference material corresponding to the cluster, the reference material contains a plurality of reference characteristic values respectively related to the transaction characteristic values; the terminal electronic device sends an analysis request related to one of the target of the customers to the server Server unit; when the server unit receives the analysis request, the server unit transmits an analysis result to the terminal electronic device, and the analysis result includes at least part of the transaction characteristic values related to the transaction characteristic data of the target customer At least part of the reference data corresponding to the cluster to which the target customer's transaction characteristic data belongs Consider the characteristic value and the cluster attribute description related to the cluster definition data that the transaction characteristic data related to the target customer meets; and when the terminal electronic device receives the analysis result, the terminal electronic device displays the analysis result . 如請求項5所述的智能分群分析方法,其中,該伺服器單元還儲存有多個分別相關於多個金融商品的金融商品歷史資料,該方法還包含:該伺服器單元根據該等金融商品歷史資料產生多個分別相關於該等金融商品的商品特徵資料,該等商品特徵資料的每一者包含多個商品特徵值;該終端電子裝置傳送一相關於該目標客戶的建議請求給該伺服器單元,該建議請求包含至少一相關於該分析結果的該等交易特徵值其中一者之目標特徵值;當該伺服器單元接收到該建議請求,該伺服器單元根據該等商品特徵資料、該至少一目標特徵值及相關於該目標客戶的交易特徵資料,產生一相關於該等金融商品其中至少一者的商品清單,並將該商品清單傳送給該終端電子裝置;及當該終端電子裝置接收到該商品清單,該終端電子裝置顯示該商品清單。The intelligent cluster analysis method according to claim 5, wherein the server unit further stores a plurality of financial commodity historical data respectively related to a plurality of financial commodities, and the method further includes: the server unit according to the financial commodities The historical data generates a plurality of product characteristic data related to the financial commodities, each of which includes a plurality of product characteristic values; the terminal electronic device sends a suggestion request related to the target customer to the server Server unit, the suggestion request includes at least one target feature value of one of the transaction feature values related to the analysis result; when the server unit receives the suggestion request, the server unit according to the product feature data, The at least one target characteristic value and transaction characteristic data related to the target customer, generating a product list related to at least one of the financial commodities, and transmitting the product list to the terminal electronic device; and when the terminal electronic device The device receives the product list, and the terminal electronic device displays the product list. 如請求項5所述的智能分群分析方法,其中,該預定分群演算法為k-平均演算法。The intelligent cluster analysis method according to claim 5, wherein the predetermined cluster algorithm is a k-average algorithm. 如請求項5所述的智能分群分析方法,其中,該終端電子裝置是以雷達圖的形式顯示該分析結果。The intelligent cluster analysis method according to claim 5, wherein the terminal electronic device displays the analysis result in the form of a radar chart.
TW106131682A 2017-09-15 2017-09-15 Intelligent cluster suggestion system and method TWI677843B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
TW106131682A TWI677843B (en) 2017-09-15 2017-09-15 Intelligent cluster suggestion system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
TW106131682A TWI677843B (en) 2017-09-15 2017-09-15 Intelligent cluster suggestion system and method

Publications (2)

Publication Number Publication Date
TW201915899A TW201915899A (en) 2019-04-16
TWI677843B true TWI677843B (en) 2019-11-21

Family

ID=66992255

Family Applications (1)

Application Number Title Priority Date Filing Date
TW106131682A TWI677843B (en) 2017-09-15 2017-09-15 Intelligent cluster suggestion system and method

Country Status (1)

Country Link
TW (1) TWI677843B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106067094A (en) * 2016-06-14 2016-11-02 华北电力大学 A kind of dynamic assessment method and system
CN106228399A (en) * 2016-07-20 2016-12-14 福建工程学院 A kind of stock trader's customer risk preference categories method based on big data
CN106326335A (en) * 2016-07-22 2017-01-11 浪潮集团有限公司 Big data classification method based on significant attribute selection
TWM543416U (en) * 2016-12-29 2017-06-11 First Commercial Bank Co Ltd Intelligent product marketing system
TWM554605U (en) * 2017-09-15 2018-01-21 Capital Securities Corp Smart advising system based on group classification

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106067094A (en) * 2016-06-14 2016-11-02 华北电力大学 A kind of dynamic assessment method and system
CN106228399A (en) * 2016-07-20 2016-12-14 福建工程学院 A kind of stock trader's customer risk preference categories method based on big data
CN106326335A (en) * 2016-07-22 2017-01-11 浪潮集团有限公司 Big data classification method based on significant attribute selection
TWM543416U (en) * 2016-12-29 2017-06-11 First Commercial Bank Co Ltd Intelligent product marketing system
TWM554605U (en) * 2017-09-15 2018-01-21 Capital Securities Corp Smart advising system based on group classification

Also Published As

Publication number Publication date
TW201915899A (en) 2019-04-16

Similar Documents

Publication Publication Date Title
US20180288177A1 (en) Methods and systems for activity-based recommendations
US10783568B1 (en) Social merchandising system
US20190108576A1 (en) Blockchain systems and methods for procurement
JP7565415B2 (en) Inter-process communication to facilitate sell-side market making
US20210182300A1 (en) Broker chat bot
US11762819B2 (en) Clustering model analysis for big data environments
KR20200114969A (en) Apparatus and method for recommending goods based on clustering of transaction history data
US20210081840A1 (en) Using a machine learning model to determine acceptability of remedial actions for supply plan deviations
JP7407543B2 (en) Information processing method, program, and information processing device
CN110490714A (en) A kind of Gross Profit from Sales real time updating method and system
WO2022126931A1 (en) Commodity recommendation method and apparatus, computer device and medium
KR20160019077A (en) Oline commerce method and system
TWI677843B (en) Intelligent cluster suggestion system and method
TWI807142B (en) System and method for managing transaction risk
TWM554605U (en) Smart advising system based on group classification
CA3026275A1 (en) Investment management proposal system
US7702570B2 (en) System and method for providing an operator interface for a radar display of market data
KR102461945B1 (en) Real-time sales appraisal transaction system to mutually mapping and managing multiple sales
US11853964B1 (en) System and method for smart inventory capture
KR102627118B1 (en) Server for prviding personalization services of stock information and method thereof
US20240144303A1 (en) Optimization apparatus, optimization method, and non-transitory computer readable medium storing optimization program
US20240354858A1 (en) Integrated portfolio rebalancing system with ai-assisted recommendations and scalable features for investment applications
TWI666605B (en) Matching system and management method of tariff quota allocation
US20160335631A1 (en) Method and system of self-imposing a financial penalty for purchasing comestibles containing nutritional content not in compliance with user pre-defined nutritional content guidelines
US20190304017A1 (en) Compound-price calculating system and method thereof

Legal Events

Date Code Title Description
MM4A Annulment or lapse of patent due to non-payment of fees